Main

Monoallelic microdeletion of chromosome 22q11.2 (del.22q11) is the most common human chromosomal deletion and results in the 22q11.2 deletion syndrome (22q11DS), with an estimated frequency of 1 in 4000–5000 births (1). 22q11DS encompasses three independently established syndromes, including DiGeorge syndrome, velo-cardio-facial syndrome, and conotruncal anomaly face syndrome, and manifests a variety of features, including craniofacial and cardiovascular defects (CVD) and behavioral and psychiatric disorders (2–4). 22q11DS is associated with specific types of CVD that are characterized as malformations of the conotruncus (outflow tract of the heart) or the aortic arch, including persistent truncus arteriosus, tetralogy of Fallot, and interrupted aortic arch, type B (5). The craniofacial defects and CVD in 22q11DS result from defects of embryonic structures populated by migrating cranial and cardiac neural crest cells (NCC) that originate from the caudal hindbrain (6, 7). Ablation of cardiac NCC in chick embryos results in persistent truncus arteriosus, aortic arch malformations, and hypoplasia of thymus and parathyroid glands similar to that observed in patients with 22q11DS (6, 7). This suggests that studies on candidate genes in the 22q11.2 region that are involved in the development of cardiac NCC may provide insight into the molecular mechanisms underlying the 22q11DS and its associated CVD.

More than 90% of patients with 22q11DS have a similar del.22q11, commonly spanning 1.5–3 megabases in size (8). More than 30 genes have been identified in the deleted region. Recent studies show that haploinsufficiency of Tbx1, one of the genes in this region, is likely to be the major determinant of aortic arch defects in patients with 22q11DS (9–11). However, no single gene mutation, including that of TBX1, that results in the 22q11DS phenotype has yet been detected (12, 13). Our screen for mouse genes dependent on dHAND, a transcription factor implicated in NCC development, previously identified Ufd1l, whose human ortholog had been mapped to 22q11.2 (14, 15). UFD1L encodes a highly conserved protein from yeast to human that is involved in the degradation of ubiquitinated proteins. Mouse Ufd1l was expressed in many embryonic tissues affected in patients with 22q11DS, including the pharyngeal arches, palatal precursors, the frontonasal region, telencephalon, and the cardiac outflow tract. Human UFD1L was deleted in >180 patients studied with 22q11DS (15). A smaller deletion of exon 1 to 3 of UFD1L and its neighboring gene, CDC45, which encodes a putative cell cycle regulating protein, was found in one individual with the phenotype of 22q11DS. These results suggest that UFD1L may be implicated in some features of 22q11DS, although the very early lethality of mice that lack Ufd1l precludes an understanding of the role of UFD1L in cardiac NCC development (16).

The avian system is well established to study NCC development, as gene constructs can be specifically introduced into the chick NCC lineage using a replication-incompetent retrovirus system (17). To take advantage of the avian system, we cloned the chick ortholog of UFD1L and performed expression and functional analyses of chick Ufd1l during embryogenesis, focusing on cardiac NCC development. Chick Ufd1l was expressed in a variety of developing tissues with enhanced expression in the cardiac outflow tract and neural crest-derived mesenchyme of pharyngeal arches and aortic arch arteries, similar to mouse Ufd1l. Attenuation of chick Ufd1l expression in premigratory cardiac NCC resulted in an increased number of conotruncal septation defects. These data suggest that Ufd1l may play a role in the development of cardiac NCC that contribute to conotruncal septation.

METHODS

Cloning of chick Ufd1l.

A stage 12–15 chick embryo library in lambdaZAPII was screened according to the manufacturer's instructions (Stratagene), with chick Ufd1l partial cDNA (741bp) that was obtained from Hamburger and Hamilton (HH) stage 12 chick total RNA by reverse-transcribed PCR. Approximately 1 μg of total RNA was reverse transcribed using M-MuLV transcriptase (Roche). One microliter of the first-strand cDNA produced was used as template for PCR reactions using specific oligonucleotide amplimers (5′-GGAGAAAGGAGGGAAGATAA-3′, 5′-GCTTCATCCTCTTCAACCTT-3′) derived from mouse Ufd1l cDNA. 5′ rapid amplification of cDNA ends was then performed to obtain full-length cDNA according to manufacture's instructions (Clontech).

Northern analysis.

Total RNA from several chick tissues at embryonic day 18 was electrophoresed, transferred to a nylon membrane, and hybridized with a 32P-labeled full-length chick Ufd1l cDNA probe. Hybridization was performed in Rapid-hyb buffer (Amersham) at 65°C overnight and washed with 2 × salt-sodium chloride/0.1% SDS at room temperature for 10 min and 1x salt-sodium phosphate-EDTA/1% SDS at 65°C for 15 min. Nylon membrane was exposed to x-ray film (Kodak).

Collection of chick embryos and primary NCC culture experiments.

Fertilized eggs were incubated at 37°C to obtain chick embryos of HH stage 10–26. The cranial neural fold from the otic vesicle to somite 3 of stage 9–11 chick embryos was dissected and treated with 0.5 mg/mL collagenase type II in PBS at 37°C for 2–3 min to remove the mesenchyme surrounding the neural tube. The isolated neural tube was washed in Dulbecco's modified Eagle's medium including 10% fetal bovine serum, then placed on gelatin-coated culture plates. The dissected neural tubes were cultured at 37°C for 36–48 h in Dulbecco's modified Eagle's medium including 10% fetal bovine serum to allow NCC to migrate from the neural tubes onto culture plates. NCC were fixed with 4% paraformaldehyde/PBS. In situ hybridization and immunohistochemistry were subsequently performed. All experimental procedures were performed according to the current local guidelines for animal experiments in the University of Texas Southwestern Medical Center and Leiden University Medical Center.

RNA in situ hybridization.

RNA in situ hybridization of chick embryos and cultured NCC was performed as previously described (18) with digoxigenin-labeled Ufd1l riboprobes synthesized from a 741-bp fragment of chick Ufd1l cDNA. Briefly, chick embryos of varying stages were isolated, and the pericardium was removed. Embryos were fixed in 4% paraformaldehyde/PBS at 4°C overnight. Embryos were treated with proteinase K (30 μg/mL) for 3–20 min at room temperature depending on their stage of development. Prehybridization was performed at 61°C for 3 h in solution containing 50% formamide, 4Ă— salt-sodium chloride, 1Ă— Denhardt's solution, 500 μg/mL salmon sperm DNA, 250 μg/mL yeast tRNA, and 10% dextran sulfate. Hybridization was performed in the same solution for 18 h. After a series of washes, embryos were treated in 2% blocking solution for 2 h and incubated with antidigoxigenin antibody for 1 h at room temperature. Color reaction was performed at room temperature in substrate color reaction mixture for 12–18 h and terminated by fixing embryos in 4% paraformaldehyde/0.1% glutaraldehyde.

Radioactive in situ hybridization was performed using 35S-labeled antisense riboprobes synthesized with T7 or SP6 RNA polymerase (MAXIscript, Ambion Inc.) from chicken dHAND or neuropilin-1 full-length cDNA, respectively. Paraffin-embedded coronal sections of HH stage 18 control embryos and antisense Ufd1l-infected embryos were hybridized with probes as previously described (18). Plasmids for dHAND and neuropilin-1 riboprobe are available upon request.

Fixed NCC were treated with proteinase K (30 μg/mL) for 5 min at 37°C. After postfixation with 4% paraformaldehyde/PBS, cells were dehydrated in methanol. In situ hybridization was performed at 61°C for 18 h as described above.

Immunocytochemistry.

For immunocytochemistry, fixed and dehydrated NCC were incubated with 3% H2O2 for 30 min at room temperature. After rehydration, cells were incubated in blocking solution [3% BSA, 0.1% Igepal (Sigma Chemical Co.) in PBS] at room temperature for 30 min and then incubated in primary antibody, anti–HNK-1/N-CAM (Sigma Chemical Co.; 1:60), at room temperature for 60 min. After several washes in PBS, conjugated secondary antibody [anti-mouse IgM (1:50)] was applied to the cells and incubated at room temperature for 60 min. The cells were then incubated in developing solution [0.03% 3-3′-diaminobenzidine tetrahydrochloride (Sigma Chemical Co.) -0.1% TritonX in PBS] at room temperature for 10 min. Color reaction was performed in 0.03% H2O2 at room temperature for 10 min. After several washes in PBS, cells were fixed with 4% paraformaldehyde/0.1% glutaraldehyde for 12–24 h.

In vitro transcription and translation.

The full-length chick Ufd1l cDNA was cloned in the sense and antisense direction into pBluescript plasmids. Mouse endothelin receptor A (ETA) was also cloned in antisense direction into pBluescript plasmids as a control. All plasmids were linearized with Sma I for in vitro transcription and translation reactions using TNTcoupled Reticulocyte Lysate System (Promega) with T7 RNA polymerase. One microgram of sense cUfd1l plasmid and antisense cUfd1l or ETA plasmid were mixed together in the same tube for transcription and translation. A sense cUfd1l plasmid and pBluescript plasmid mixture was performed simultaneously as a control. Translated products were labeled with [35S] methionine (Amersham) and determined by SDS-PAGE and autoradiography. The signals were quantified using ImageQuant software (Molecular Dynamics).

Retrovirus infection of chick embryos.

The full-length chick Ufd1l cDNA was cloned in the sense and antisense direction into the proretroviral pCXIZ vector (gift from Dr. Takashi Mikawa). All retroviral constructs were stably transfected to the D17.2G helper cell line for virus production. In vivo infection of premigrating NCC in chick embryos was performed at HH stage 9–11 as described previously (19, 20). The embryos were allowed to develop in ovo until HH stage 31–32 and then were harvested and fixed with 4% paraformaldehyde for histologic analysis. HH stage 18 embryos were also harvested for section in situ hybridization using dHAND and neuropilin-1 riboprobes.

In vitro infection was performed on neural tube dissected from HH stage 9–11 as described above. After NCC migration, β-galactosidase activity was detected by using X-gal as described previously (21).

Histology.

Embryos were embedded in paraffin, and serial transverse sections were obtained. The transverse sections of embryos after retrovirus infection or control embryos were stained with hematoxylin and eosin.

RESULTS

Chick Ufd1l cDNA.

The largest cDNA obtained of chick Ufd1l by library screening followed by 5′ rapid amplification of cDNA ends contained 1026 nucleotides. Chick Ufd1l was predicted to encode a 307–amino acid protein and was highly conserved with its murine (94% identical, 96% similar) and human (94% identical, 96% similar) counterparts (Fig. 1A). Although the previously cloned human UFD1L protein (as translated from cDNA, GenBank accession number U64444) (14) showed a tandem repetition of a 36–amino acid domain that was present in the yeast, chicken, and mouse as a single copy, the protein sequences predicted from all human UFD1L cDNAs that we cloned (GenBank accession number AY101594) did not contain such a domain and were more similar to those of other species. Chick Ufd1l was also 36% identical and 50% similar to yeast Ufd1. Nucleotide and protein sequences of chick Ufd1l are available at NCBI (GenBank accession number AF407671).

Figure 1
figure 1

Sequence and expression of chick Ufd1l. A, Sequence alignment of full-length chick Ufd1l with mouse, human, and yeast Ufd1l. Amino acid residues that are identical or similar are shaded in black or gray, respectively. Chick Ufd1l encodes a 307–amino acid protein and is highly conserved with its murine (94% identical, 96% similar) and human (94% identical, 96% similar) counterparts. It was 36% identical and 50% similar to yeast Ufd1. B, Northern blot analysis of chick Ufd1l in embryonic day 18 tissues. mRNA transcripts (1.2, 1.8, and 3.0 kb) were present in all tissues examined. The expression levels of chick Ufd1l in brain, heart, skeletal muscle, and thymus were higher than in other tissues.

Full size image

Expression pattern of chick Ufd1l.

Northern blot analysis of total RNA from several chick tissues at embryonic day 18 was performed using full-length chick Ufd1l cDNA as a probe. mRNA transcripts (1.2, 1.8, and 3.0 kb) were present in all tissues examined and possibly represented alternatively spliced forms (Fig. 1B). The expression level of chick Ufd1l in each tissue was variable, with expression in brain, heart, skeletal muscle, and thymus relatively higher than in other tissues.

To determine the expression pattern of chick Ufd1l during embryogenesis, in situ hybridization analysis was performed. Chick Ufd1l was detectable by stage 7 in the neural plate (data not shown). At stage 10, as the neural tube and somites formed, expression throughout the neural tube was apparent (Fig. 2A). At later stages, chick Ufd1l was expressed in many tissues, with marked expression in the head, pharyngeal arches, frontonasal region, neural tube, and limb buds (Fig. 2, B and C). Whole-mount in situ hybridization on the heart dissected from embryos at stage 26, when conotruncal septation takes place, demonstrated strong expression of chick Ufd1l in the conotruncal region of the heart and in emerging aortic arch arteries (Fig. 2D). The expression of chick Ufd1l in the neural tube and mesenchymal cells along the pharyngeal arch arteries was detected in transverse sections through chick embryos after whole-mount in situ hybridization (Fig. 2, E and F). The overall embryonic expression pattern of chick Ufd1l was similar to that of mouse (14, 15) and consistent with some of the clinical features of 22q11DS.

Figure 2
figure 2

Expression of chick Ufd1l during embryogenesis. In situ hybridization of chick Ufd1l on stage 10, 16, or 21 chick embryos (A–C) and on heart of stage 26 embryo (D). At stage 10, the expression throughout the neural tube (nt) was apparent (A). At later stages, chick Ufd1l expression was enhanced in the head (h), pharyngeal arches (pa), frontonasal region (fn), neural tube (nt, white arrow heads), and limb buds (lb) (B and C). D, Strong expression of chick Ufd1l in the outflow tract (ot) of the heart and arch arteries (white arrows). Transverse section of embryo in A and B after whole-mount in situ hybridization (E and F, respectively) is shown at levels indicated. The expression was observed in the neural tube (nt), cranial mesenchyme (*), aortic arch arteries (black arrow heads), and outflow tract (ot) (E and F). ph, pharynx; ht, heart.

Full size image

To determine whether chick Ufd1l was expressed in migrating NCC, we examined primary chick NCC cultures. NCC that had migrated off chick neural tube in culture exhibited normal expression of the avian NCC marker HNK-1 by immunocytochemistry (Fig. 3A). In situ hybridization in the cultured NCC demonstrated chick Ufd1l expression in the migrating NCC, consistent with the result of whole-mount in situ hybridization analysis on embryos (Fig. 3, B and C).

Figure 3
figure 3

Chick Ufd1l is expressed in primary neural crest culture. A, HNK-1, as a NCC marker, was detected by immunocytochemistry in primary neural crest cultured cells. B, Chick Ufd1l transcripts were detected by antisense mRNA in situ hybridization in primary cultured cells. C, In situ hybridization with sense Ufd1l riboprobe as control.

Full size image

Functional attenuation of chick Ufd1l by retroviral infection.

To investigate the functional role of chick Ufd1l in cardiac NCC, we used a well-established gene introduction system in chick NCC using a replication-incompetent retrovirus (17). To study loss of function effects in vivo, we prepared a Ufd1l antisense retroviral vector that would generate Ufd1l antisense transcripts. To determine first whether the Ufd1l antisense transcripts might interfere with normal translation of Ufd1l sense transcripts, we performed in vitro transcription and translation experiments. As shown in Figure 4, the protein expression of chick Ufd1l was reduced ∼50% by expression of the antisense chick Ufd1l transcript. The expression of control (ETA) antisense transcripts did not affect the protein expression of chick Ufd1l, suggesting that the effect of antisense chick Ufd1l introduction was specific for the reduction of chick Ufd1l protein expression. To determine whether cells infected with the Ufd1l antisense vector had a nonspecific survival defect, we performed primary neural crest culture from neural tube infected with the Ufd1l antisense retrovirus. Retrovirus-infected cells were marked in blue by β-galactosidase staining, and we found that their survival was unaffected by Ufd1l antisense transcripts (Fig. 4B).

Figure 4
figure 4

Antisense chick Ufd1l transcripts can attenuate chick Ufd1l protein expression. A, Quantitative analysis of in vitro transcription and translation reveals a 54% reduction in Ufd1l protein when Ufd1l sense (S) and antisense (AS) transcripts were co-incubated compared with S alone or S with a control (C) antisense (ETA). The expression of ETA antisense transcripts did not affect protein expression of chick Ufd1l. B, Primary NCC culture after neural tube infection with retrovirus containing the AS Ufd1l expression construct linked to β-galactosidase. Survival and migration of infected cells (blue) were unaffected.

Full size image

For determining the functional role played by chick Ufd1l in cardiac NCC, the rostral neural tube was infected with retrovirus containing the antisense Ufd1l expression construct before NCC migration. By examining the hearts of a number of embryos at HH stage 31–32, 6 d after the retroviral infection, we found a significant (p < 0.05) increase in septal defects (nine of 12) compared with control embryos (two of 12 and zero of 12, respectively) that were infected with retrovirus containing an empty vector (Fig. 5, A–F). Five of 12 experimental embryos had defects of the upper portion of the ventricular septum (VSD;Fig. 5, B and E). An additional four of 12 embryos had VSD and defects in septation of the aorta and pulmonary artery (conotruncal septal defects;Fig. 5, C and F). This combination was never observed in control embryos (p < 0.05), whereas isolated VSD were observed in two of 12 control embryos. Examination of the aortic arch in all embryos failed to detect any defects at this stage. Results are summarized in Table 1.

Figure 5
figure 5

Attenuation of Ufd1l in chick neural crest results in conotruncal defects. Transverse sections of the stage 31–32 chick hearts at the level of the conotruncal septum (A–C) or ventricular septum (D–F) in control embryos (A and D) or retrovirus Ufd1l antisense (AS) infected embryos (B, C, E, and F). G, Schematic indicating level of section. Ventricular septal defects (arrow) were observed in AS-infected embryos (B and E). Conotruncal septal defects (arrowhead in C) accompanied with VSD (arrowhead in F) were observed in AS-infected embryos. Radioactive RNA in situ hybridization on coronal sections of HH stage 18 embryos revealed similar expression of dHAND, a marker of neural crest–derived mesenchyme, in control (H) and AS-infected (I) embryos. Ao, aorta; PA, pulmonary artery; lv, left ventricle; lvot, left ventricular outflow tract; rv, right ventricle; rvot, right ventricular outflow tract; ph, pharyngeal arch.

Full size image
Table 1 cUfd11 AS retrovirus-infected embryos display cardiac outflow tract defects
Full size table

To investigate whether NCC were specified and migrated in antisense Ufd1-infected embryos, we examined two molecular markers, dHAND and neuropilin-1, both expressed in the NCC-derived mesenchyme of pharyngeal arches. Expression of dHAND (Fig. 5, H and I) and neuropilin-1 (data not shown) was relatively normal in the pharyngeal arches of HH stage 18 experimental embryos compared with controls. This result suggests that NCC become specified, migrate, and differentiate into mesenchymal cells during early pharyngeal development of antisense Ufd1l-infected embryos. Taken together, these data indicate that functional attenuation of Ufd1l in cardiac NCC leads to an increased incidence of conotruncal septal defects in chick embryos and these defects do not result from nonspecific cell death before NCC migration.

DISCUSSION

Cross-species analysis is sometimes useful in understanding potential causes of human disease. Mice homozygous for deletion of Ufd1l are early embryonic lethal, whereas mice heterozygous for a Ufd1l mutation are normal, precluding an examination of whether Ufd1l plays a role in cardiac NCC development (16). Antisense nucleotides designed to hybridize to specific mRNA sequences have been used to attenuate the function of numerous genes both in vitro and in vivo, including developmentally regulated genes in chick embryos (22–24). By using a replication-incompetent retrovirus that has proved beneficial in studying the avian NCC lineage, we were able to introduce antisense Ufd1l transcripts into chick cardiac NCC over long time spans and examine the role of Ufd1l in cardiac NCC in vivo.

Although functional attenuation of Ufd1l in NCC lineage resulted in an increased number of conotruncal septal defects in chick embryos, there was phenotypic variation. It is possible that the phenotypic variation may reflect the gene dosage effect of Ufd1l resulting from a variable efficiency of retroviral infection in each chick embryo. If this is the case, then conotruncal septation requires Ufd1l in a dose-sensitive manner in this chick model, although haploinsufficiency of Ufd1l alone is not sufficient for the septal defects in the mouse model. Alternatively, it may reflect allelic variability, variable penetrance caused by environmental factors, or stochastic events during embryonic development similar to that proposed in patients with 22q11DS. We cannot distinguish between these possibilities as it is not possible to examine effectively the level of attenuation of Ufd1l protein in premigratory NCC of embryos studied later in development for defects in cardiogenesis.

Recent genomic manipulations in mice have provided insight into potential roles of candidate genes for 22q11DS. Mice harboring a hemizygous deletion of 18 genes, including Ufd1l, that are deleted in patients with 22q11DS (Df1/+) were generated and exhibit aortic arch malformations in ∼30% of cases (16). Further analyses determined that haploinsufficiency of Tbx1 was responsible for aortic arch malformations in Df1/+ mice (10). Mice homozygous for the Tbx1 mutation displayed craniofacial and cardiovascular defects similar to 22q11DS (9–11), suggesting that TBX1 may be a major determinant of the phenotype of 22q11DS. However, TBX1 mutations have not been identified in patients who have the 22q11 deletion phenotype and do not have del.22q11 (10, 13). In addition, Df1/+ mice have broader phenotypic features that have not been observed in mice heterozygous for a Tbx1 mutation, including conotruncal and ventricular septal defect, hypoplasia of thymus and parathyroid glands, and signs of behavioral and psychiatric disorder (16, 25, 26). These results indicate that haploinsufficiency of Tbx1 and additional genes in the Df1 deletion region together contribute to the entire phenotype of Df1/+ mice. Our data in the present study suggest a role for Ufd1l in conotruncal development and thus support a possibility that Ufd1l may contribute to conotruncal and ventricular septal defects as a modifier gene in Df1/+ mice. Ufd1l might also contribute to the hypoplasia of the thymus/parathyroid glands and behavioral/psychiatric phenotype of Df1/+ mice, because it is expressed in the developing pharyngeal structures and brain. The association of a promoter polymorphism of UFD1L with schizophrenia susceptibility has recently been reported (27). Definitive evaluation of Ufd1l's role in particular tissues awaits tissue-specific disruption of Ufd1l in mice, currently being performed in our laboratory.

Although we determined that Ufd1l plays a role in cardiac NCC, cellular functions of Ufd1l in avian or mammalian cells remain to be studied. We speculate that Ufd1l is required for postmigratory differentiation of cardiac NCC in the conotruncal region because 1) the infection of retrovirus containing antisense Ufd1l did not affect NCC survival and migration in primary NCC;2) Ufd1l was expressed at high levels in the conotruncal region; and 3) functional attenuation of Ufd1l resulted in conotruncal septation defects, although the phenotype was milder than that observed after ablation of premigratory NCC in chick embryos. Recent biochemical studies demonstrated multiple functions of Ufd1, including the activation of membrane-bound transcription factors and transportation of proteins from the endoplasmic reticulum into the cytosol, by regulating ubiquitin/proteosome-dependent processing (28–31). Identification of pathways affected by Ufd1l in cardiac NCC will be necessary to understand the mechanism through which Ufd1l regulates NCC development.